This replication archive contains the files and data necessary to replicate results from Matthew Blackwell (2013), "A Selection Bias Approach to Sensitivity Analysis for Causal Effects." Included files are:

* causalsens_0.1.tar.gz - R package source for the implementation of the sensitivity analysis for the first two (static) illustrations. Updated versions of the package may continue to work, but this version can always be installed to reproduce the exact results of the paper. You can install this by package by running the following command in R: 

    install.packages("causalsens_0.1.tar.gz", repos = NULL, type = "source")
    
* figure1.R - R code to load the LaLonde data, run the relevant analyses, and generate Figure 1.
* psid_controls.dta - Dataset containing the non-experimental controls used in the LaLonde example.
* figure2.R - R code to load the Boyd, Epstein, and Martin (2010) data, run the relevant analyses, and generate Figure 2.
* genderjudging.dta - Dataset from Boy, Epstein, and Martin (2010). Original data and study availabe here: http://epstein.usc.edu/research/genderjudging.html
* figure3.R - R code to load campaign negativity and turnout data, run analyses, and generate Figure 3. Note that this file does not use the causalsens package. 
* dem.panel.RData - R data file containing the campaign negativity data. 
* panel-utils.R - set of functions used to produce Figure 3. 
* README.md - this file. 

